Three-Dimensional Reconstruction of Targets Based on Multi-angle High-Resolution Range Profiles

Ahstract-The two main features of the multi-angle netted radar data acquisition are the echo information diversity and spatial sparsity, which brings about opportunities and challenges respectively. This paper is the first toe xploit t he potential of the multi-angle radar in targets' three-dimensional (3D) reconstruction, under the condition that multiple stations provide one-dimensional (1D) range profiles from every single observation angle. The main idea is to reconstruct all possible samples of correlation using geometric constraints, phase difference relationship and conversion of coordinates. Under the principle of random sampling consistency (RANSAC), a certain number of possible correlated samples are traversed, for reconstruct real target scattering points as well as eliminating false correlated samples. From a practical point of view, there exists random estimation deviation in terms of the platform positioning, time synchronization, platform's observation angle. Hence, error analysis is theoretically derived and simulated under a typical geometry of airborne multi-angle radar. The simulation results have verified t he effectiveness a nd performance b ound of the proposed algorithm.